Research On Application Of Financial Crisis Early Warning Model With Text Information Index

With the implementation of the "Made in China 2025" action program and the adjustment of the national industrial policy,the proportion of the manufacturing industry in the national economic aggregate has decreased,but the production capacity contained in the manufacturing industry itself is the key to the long-term development of the national economy.At present,manufacturing enterprises are faced with the complexity of market environment and the uncertainty of management risk,and may fall into financial crisis at any time.As the manufacturing industry is characterized by close upstream and downstream industry chains,once a financial crisis occurs,it will cause significant losses to stakeholders in the whole industry chain.The high-tech communication equipment manufacturing industry is characterized by large investment in research and development and rapid technology update and iteration.Enterprises in the industry are prone to financial crisis due to mistakes in development and management decisions.Therefore,the analysis of enterprise financial risk and financial crisis warning have become the focus of concern of enterprises and their stakeholders.Currently,about manufacturing enterprise financial crisis warning research work mainly focused on the basis of the data in the financial statements,to build early warning index and input to the machine learning model to forecast and analysis,this method the research perspectives focus on the financial statements of the static data,ignore the external public opinion on the influence of enterprise financial crisis,has certain limitation;The financial crisis early warning model,introduced the text information index,text information index of the use of text mining technology to build emotion,can comprehensively reflect the enterprise’s external evaluation case and relevant opinion heat,and the real-time and accurately grasp the changes of the market,can promote the timeliness and accuracy of the financial crisis warning.This thesis takes communication equipment manufacturer A company as the case study object.Firstly,the definition and causes of financial crisis,financial crisis early warning index,models and text information in the application of financial crisis early warning research literature at home and abroad are sorted out,and the theoretical basis of financial crisis early warning and the characteristics and applications of Logistic regression are described.Secondly,the limitations of Logistic regression in the application of financial crisis warning and the necessity of introducing text information index are analyzed.Thirdly,combined with the financial risk characteristics of communication equipment manufacturing enterprises,the data of 60 sample companies of similar size from 2012 to 2020 are selected,the early-warning index based on financial statement data are selected,and the text information index are constructed by using web crawler technology and text analysis technology.The text information index was introduced into the financial crisis warning model based on Logistic regression.After the multicollinearity test,K-S normal distribution test,T test and significance test of MannWhitney U index of all the warning indexes,nine warning indexes with significance for financial crisis warning were selected.A financial crisis warning model with text information index was constructed,and Hosmer-Lemeshow test was used to test the goodness of fit of the model.The test results showed that the model had good fitting effect.At the same time,the accuracy of the prediction is tested,the results show that the comprehensive prediction accuracy of the model is 81%,the model of financial crisis prediction effect is better;Then,the financial crisis warning model after introducing text information index is applied to the financial crisis warning analysis of A company.The analysis results show that the model can accurately forecast the financial crisis warning situation of A company by using the data of 3 years before the financial crisis.Finally,the thesis puts forward corresponding countermeasures and suggestions for A company’s financial crisis alarm.In this thesis,crawler technology and text analysis technology are used to construct text information index,and the text information index is introduced into the financial crisis early warning model,which provides a new perspective for the enterprise financial crisis early warning,and provides a beneficial supplement for the perfection of financial crisis early warning theory system.In addition,this thesis will introduce text information index of financial crisis early warning model applied to the actual case of enterprise financial crisis early warning analysis,in order to verify the scientific and effective introduction of index,not only help enterprises accurately analyze the characteristics of financial crisis,improve the accuracy of enterprise financial crisis early warning;It also helps enterprises to formulate relevant financial crisis prevention and coping strategies,and has strong practical guiding significance for enterprises to prevent and deal with financial crisis in time.